Dr Haris Pervaiz

In this paper, we present an analytical framework to model the sleep mode power consumption of a base station (BS) as a function of its sleep depth. The sleep depth is made up of the BS deactivation latency, actual sleep period and activation latency. Numerical results demonstrate a close match between our proposed approach and the actual sleep mode power consumption for selected BS types. As an application of our proposed approach, we analyze the optimal sleep depth of a BS, taking into consideration the increased power consumption during BS activation, which exceeds its no-load power consumption. We also consider the power consumed during BS deactivation, which also exceeds the power consumed when the actual sleep level is attained. From the results, we can observe that the average total power consumption of a BS monotonically decreases with the sleep depth as long as the ratio between the actual sleep period and the transition latency (deactivation plus reactivation latency) exceeds a certain threshold.

Nowadays, system architecture of the fifth generation (5G) cellular system is becoming of increasing interest. To reach the ambitious 5G targets, a dense base station (BS) deployment paradigm is being considered. In this case, the conventional always-on service approach may not be suitable due to the linear energy/density relationship when the BSs are always kept on. This suggests a dynamic on/off BS operation to reduce the energy consumption. However, this approach may create coverage holes and the BS activation delay in terms of hardware transition latency and software reloading could result in service disruption. To tackle these issues, we propose a predictive BS activation scheme under the control/data separation architecture (CDSA). The proposed scheme exploits user context information, network parameters, BS sleep depth and measurement databases to send timely predictive activation requests in advance before the connection is switched to the sleeping BS. An analytical model is developed and closed-form expressions are provided for the predictive activation criteria. Analytical and simulation results show that the proposed scheme achieves a high BS activation accuracy with low errors w.r.t. the optimum activation time.

IoT data analytics is underpinning numerous applications, however the task is still challenging predominantly due to heterogeneous IoT data streams, unreliable networks and ever increasing size of the data. In this context, we propose a two layer architecture for analyzing IoT data. The first layer provides a generic interface using a service oriented gateway to ingest data from multiple interfaces and IoT systems, store it in a scalable manner and analyze it in real-time to extract high-level events whereas second layer is responsible for probabilistic fusion of these high-level events. In the second layer, we extend state-ofthe- art event processing using Bayesian networks (BNs) in order to take uncertainty into account while detecting complex events. We implement our proposed solution using open source components optimized for large-scale applications. We demonstrate our solution on real-world use-case in the domain of intelligent transportation system (ITS) where we analysed traffic, weather and social media data streams from Madrid city in order to predict probability of congestion in real-time. The performance of the system is evaluated qualitatively using a web-interface where traffic administrators can provide the feedback about the quality of predictions and quantitatively using F-measure with an accuracy of over 80%.

Hybrid networks consisting of both millimeter wave (mmWave) and microwave (¼W) capabilities are strongly contested for next generation cellular communications. A similar avenue of current research is deviceto- device (D2D) communications, where users establish direct links with each other rather than using central base stations (BSs). However, a hybrid network, where D2D transmissions coexist, requires special attention in terms of efficient resource allocation. This paper investigates dynamic resource sharing between network entities in a downlink (DL) transmission scheme to maximize energy efficiency (EE) of the cellular users (CUs) served by either (¼W) macrocells or mmWave small cells, while maintaining a minimum quality-of-service (QoS) for the D2D users. To address this problem, firstly a self-adaptive power control mechanism for the D2D pairs is formulated, subject to an interference threshold for the CUs while satisfying their minimum QoS level. Subsequently, a EE optimization problem, which is aimed at maximizing the EE for both CUs and D2D pairs, has been solved. Simulation results demonstrate the effectiveness of our proposed algorithm, which studies the inherent tradeoffs between system EE, system sum rate and outage probability for various QoS levels and varying density of D2D pairs and CUs.

Ultra densification in heterogeneous networks (HetNets) and the advent of millimeter wave (mmWave) technology for fifth generation (5G) networks have led the researchers to redesign the existing resource management techniques. A salient feature of this activity is to accentuate the importance of computationally intelligent (CI) resource allocation schemes offering less complexity and overhead. This paper overviews the existing literature on resource management in mmWave-based HetNets with a special emphasis on CI techniques and further proposes frameworks that ensure quality-of-service requirements for all network entities. More specifically, HetNets with mmWavebased small cells pose different challenges as compared to an allmicrowave- based system. Similarly, various modes of small cell access policies and operations of base stations in dual mode, i.e., operating both mmWave and microwave links simultaneously, offer unique challenges to resource allocation. Furthermore, the use of multi-slope path loss models becomes inevitable for analysis owing to irregular cell patterns and blocking characteristics of mmWave communications. This paper amalgamates the unique challenges posed because of the aforementioned recent developments and proposes various CI-based techniques including game theory and optimization routines to perform efficient resource management.

We consider an idealistic scenario where the vacation (no-load) period of a typical base station (BS) is known in advance such that its vacation time can be matched with a sleep depth. The latter is the sum of the deactivation latency, actual sleep period and reactivation latency. Noting that the power consumed during the actual sleep period is a function of the deactivation latency, we derive an accurate closed-form expression for the optimal deactivation latency for deterministic BS vacation time. Further, using this expression, we derive the optimal average power consumption for the case where the vacation time follows a known distribution. Numerical results show that significant power consumption savings can be achieved in the sleep mode by selecting the optimal deactivation latency for each vacation period. Furthermore, our results also show that deactivating the BS hardware is sub-optimal for BS vacation less than a particular threshold value.

Nowadays, dense network deployment is being considered as one of the effective strategies to meet capacity and connectivity demands of the fifth generation (5G) cellular system. Among several challenges, energy consumption will be a critical consideration in the 5G era. In this direction, base station on/off operation, i.e., sleep mode, is an effective technique to mitigate the excessive energy consumption in ultra-dense cellular networks. However, current implementation of this technique is unsuitable for dynamic networks with fluctuating traffic profiles due to coverage constraints, quality-of-service requirements and hardware switching latency. In this direction, we propose an energy/load proportional approach for 5G base stations with control/data plane separation. The proposed approach depends on a multi-step sleep mode profiling, and predicts the base station vacation time in advance. Such a prediction enables selecting the best sleep mode strategy whilst minimizing the effect of base station activation/reactivation latency, resulting in significant energy saving gains.